Just noticed a rather glaring typo, so here is the edit: When I said wireless router, I actually meant using an individual computer's wireless card as a component in a mesh-network. I think the idea of transient spatial networks is rather sooncoming and futuristic in the sense that it is adaptive evolution of technology... It would be cool to build something that would allow anybody to broadcast their own wireless network in a radius from their machine, and have other people that connect do the same, so that effectively you get a mini pooled intranet of sorts.
The first asynchronous ui-morphing language that got to popularity is js! Not saying it's awesome, but as Rich Hickey would say, Javascript _reaches_. Considering that, until the next [really good] iteration of ui-affective languages and browser-integrations comes along, js is what we have, as a world of devs, to work with.
This is something I've been postponing for a long time. I'm mostly delaying it for the "dread" of having to pick and learn a GUI library.
A quick and dirty content agnostic grid based level editor.
When making some game prototypes and in Ludum Dares I find myself making grid based games. I use text files or excel to "build" my levels, but it is cumbersome and not very flexible.
What I want is:
- A grid view. Where I can "dig" cells. Either the grid is infinite or I can add/remove columns and rows.
- A generic "level content manager". I can add into the project a set of "tags" (Enemy X, item Y, trap Z, event W) with a visual representation (shapes and colors)
- I can add these previously created "content" to cells in the map.
- It exports to a generic format. JSON is the first one that comes to mind.
EXTRA: Export to Lua. With the option to add custom code bits to cells/content that is semantically checked.
Ogmo Editor by Matt Thorson (of TowerFall fame) seems to have been created to fit exactly this niche. Exports to XML but it's open-source so you could always change that.
Given how often people complain about the modern web being app-centric, despite HTML, CSS and JS being meant for a document-centered model (and JS really only existing as a gimmick to sell Netscape,) I've been wondering just how well (or poorly) the web model could work with streaming applications.
In other words, instead of hacks that replicate a UI in the browser, have something that isn't a browser but acts like it, and source code repositories at URLs instead of documents which compile and run per request, or session (if you want to "cache" them,) or whatever. This would differ from embedded applets, flash or similar in that you would be pointing to code rather than streaming and running binaries, which would enable the user to view or possibly modify the source code before running.
Probably a terrible idea for numerous reasons, or it more or less already exists and I just don't know about it (which is not unlikely) but it might be interesting to see how it might work, and it would be far closer to the FOSS ideal than the web, for which you can't actually see or modify anything other than the result of requests.
Not much different, but the basic idea is to not be limited to a particular runtime or language, and to get away from transpiling other languages to javascript. Essentially, serving containers or emulators with native code, rather than documents and JS.
A client-side transpiler from something like WPF to HTML.
I want native, rich binding of data in whatever directions necessary and a model that makes sense. I don't want any bleed from HTML oddities, or even any access to the HTML layer. Abstract that away from me entirely with something that makes sense. Let me publish packages so I can deploy UI components in a nice tiny bundle. I'm fine with Javascript being the code layer, but give me a modern render layer, even if under the hood it's built on quirky HTML.
I feel like every library out there is a half-attempt at this, but every single one ends up being "HTML, but with data binding". I don't want "HTML with something". I want a modern presentation layer, using modern concepts.
Angular kind of does a decent job at binding, but I still need to handle weird corner cases in HTML and hack things together. Polymer is further from HTML, but the binding syntax is pretty gimped, and you end up using a bunch of div's anyway to get things just right.
Maybe this exists, but my Google-fu is to weak to find it.
Of the things I wish existed, the thing that's most well defined would be a replacement set of commands to the Git Porcelain commands that output structured data (e.g. JSON) instead of plaintext.
20 comments
[ 0.20 ms ] story [ 81.0 ms ] thread+ multi-platform software that let me use my wireless router as a wireless mesh network... is possible?
+ hydroponic year-round grocery store & restaurant
+ screen-based multiplayer music games
+ plant identification software for phones
https://www.npmjs.com/package/timemachine
:)
Very naive example:
- git add .
- git commit -m "Some commit message"
- git push
Becomes: newcmd . "Some commit message"
A quick and dirty content agnostic grid based level editor.
When making some game prototypes and in Ludum Dares I find myself making grid based games. I use text files or excel to "build" my levels, but it is cumbersome and not very flexible.
What I want is:
- A grid view. Where I can "dig" cells. Either the grid is infinite or I can add/remove columns and rows.
- A generic "level content manager". I can add into the project a set of "tags" (Enemy X, item Y, trap Z, event W) with a visual representation (shapes and colors)
- I can add these previously created "content" to cells in the map.
- It exports to a generic format. JSON is the first one that comes to mind.
EXTRA: Export to Lua. With the option to add custom code bits to cells/content that is semantically checked.
http://www.ogmoeditor.com/
In other words, instead of hacks that replicate a UI in the browser, have something that isn't a browser but acts like it, and source code repositories at URLs instead of documents which compile and run per request, or session (if you want to "cache" them,) or whatever. This would differ from embedded applets, flash or similar in that you would be pointing to code rather than streaming and running binaries, which would enable the user to view or possibly modify the source code before running.
Probably a terrible idea for numerous reasons, or it more or less already exists and I just don't know about it (which is not unlikely) but it might be interesting to see how it might work, and it would be far closer to the FOSS ideal than the web, for which you can't actually see or modify anything other than the result of requests.
I want native, rich binding of data in whatever directions necessary and a model that makes sense. I don't want any bleed from HTML oddities, or even any access to the HTML layer. Abstract that away from me entirely with something that makes sense. Let me publish packages so I can deploy UI components in a nice tiny bundle. I'm fine with Javascript being the code layer, but give me a modern render layer, even if under the hood it's built on quirky HTML.
I feel like every library out there is a half-attempt at this, but every single one ends up being "HTML, but with data binding". I don't want "HTML with something". I want a modern presentation layer, using modern concepts.
Angular kind of does a decent job at binding, but I still need to handle weird corner cases in HTML and hack things together. Polymer is further from HTML, but the binding syntax is pretty gimped, and you end up using a bunch of div's anyway to get things just right.
Maybe this exists, but my Google-fu is to weak to find it.